Data Tree Checkpoint and Restoration System and Method

1. A method of managing data in a data storage system comprising: share-latching a first node of a data tree to be copied into persistent storage; copying the first node that is share-latched into persistent storage; determining if there is a sibling second node linked to the first node; and in response to the presence of a sibling second node linked to the first node, following a link between the first node that was copied and the sibling second node, share-latching the sibling second node, unlatching the first node that was copied, and copying the sibling second node into persistent storage.

2. The method of claim 1 , further comprising determining if there is a sibling third node linked to the right of the sibling second node, and in response to a sibling third node linked to the sibling second node, following a link between the sibling second node and the sibling third node, share-latching the sibling third node, unlatching the sibling second node, and copying the sibling third node into persistent storage.

3. The method of claim 1 , further comprising, in response to there being no sibling node linked to a node that was copied, unlatching the node that was copied.

4. The method of claim 3 , wherein in response to there being no sibling node linked to a node that was copied, the method further comprises determining whether the method should proceed to copy an upper level of nodes into persistent storage.

5. The method of claim 4 , wherein in response to determining that the method should proceed to copy an upper level of nodes into persistent storage, copying non-leaf nodes into persistent storage.

6. The method of claim 5 , wherein pointers to nodes are copied into persistent storage.

7. The method of claim 1 , wherein each node that is copied into persistent storage contains two or more records.

8. The method of claim 1 , wherein only leaf nodes containing records are copied into persistent storage.

9. The method of claim 1 , further comprising restoring a data tree structure by copying the nodes in persistent storage into memory.

10. The method of claim 9 , wherein restoring the data tree structure comprises identifying a most recent complete checkpoint file.

11. The method of claim 10 , wherein restoring the data tree structure further comprises reading the nodes in the most recent complete checkpoint file from persistent storage and copying the nodes read from the most recent complete checkpoint file into memory.

12. The method of claim 11 , wherein restoring the data tree structure further comprises copying the nodes from the most recent complete checkpoint file in persistent storage concurrently into memory.

13. The method of claim 11 , wherein restoring the data tree structure comprises updating the nodes copied from the most recent compete checkpoint file based upon data written since the most recent complete checkpoint file finished.

14. The method of claim 13 , wherein restoring the data tree structure comprises updating the nodes read from the most recent complete checkpoint file to include data from records written and stored in a record log.

15. The method of claim 11 , further comprising setting pointers for the nodes copied from persistent storage into memory and the updated nodes.

16. A computer program product, the computer program product embodied on one or more computer-readable media and comprising programming instructions that when executed cause a processor to: share-latch a first node of a data tree to be copied into persistent storage; copy the first node that is share-latched into persistent storage; determine if there is a sibling second node linked to the right of the first node; and in response to a sibling second node linked to the right of the first node, follow a link between the first node that was copied and the sibling second node, share-latch the sibling second node, unlatch the first node that was copied, and copy the sibling second node into persistent storage.

17. The computer program product of claim 16 , wherein the computer program product further comprises programming instructions that when executed cause a processor to: determine if there is a sibling third node linked to the right of the sibling second node, and in response to a sibling third node linked to the right of the sibling second node: follow a link between the sibling second node and the sibling third node, share-latch the sibling third node, unlatch the sibling second node, and copy the sibling third node into persistent storage.

18. The computer program product of claim 17 , wherein the computer program product further comprises programming instructions that when executed cause a processor to copy only nodes containing two or more records into persistent storage.

19. The computer program product of claim 18 , wherein the computer program product further comprises programming instructions that when executed cause a processor to: restore a data tree structure by: identifying the most recent complete checkpoint file; reading the nodes in the most recent complete checkpoint file from persistent storage; and copying the nodes read from the most recent complete checkpoint file from persistent storage concurrently and directly into memory without buffering.

20. The computer program product of claim 19 , wherein the computer program product further comprises programming instructions that when executed cause a processor to: restore the data tree structure by updating the checkpoint nodes read from the most recent complete checkpoint file to include data from records written and stored in a record log.